The Super Mario Bros movie earlier this year broke box office records and introduced a new generation to a host of the franchise’s iconic characters. But one Mario character that wasn’t even in the megahit is somehow the perfect avatar for the 2023 zeitgeist, where artificial intelligence has suddenly arrived on the scene: Waluigi, of course. See, Mario has a brother, Luigi, and both of them have evil counterparts, the creatively named Wario and Waluigi (because Wario has Mario’s “M” turned the other way on his ever-present hat, naturally). Likely inspired by the Superman villain Bizarro, who since 1958 has been the evil mirror image of Superman from another dimension, the “Waluigi effect” has become a stand-in for a certain type of interaction with A.I. You can probably see where this is going …
The “Waluigi effect” theory goes that it becomes easier for A.I. systems fed with seemingly benign training data to go rogue and blurt out the opposite of what users were looking for, creating a potentially malignant alter-ego. Basically, the more information we trust to A.I., the higher the chances an algorithm can warp its knowledge for an unintended purpose. It’s already happened several times, like when Microsoft’s Bing A.I. threatened users and called them liars when it was clearly wrong, or when ChatGPT was tricked into adopting a rash new persona that included being a Hitler apologist.
